Explore the influences of AR supported simulation on mutual engagement of social interaction in face to face collaborative learning for physics

EXPLORE THE INFLUENCES OF AR-SUPPORTED SIMULATION ON MUTUAL ENGAGEMENT OF SOCIAL INTERACTION IN FACE-TO-FACE COLLABORATIVE LEARNING FOR PHYSICS LI NAI (B.A., TSINGHUA UNIVERSITY) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF ARTS DEPARTMENT OF COMMUNICATIONS AND NEW MEDIA NATIONAL UNIVERSITY OF SINGAPORE 2012 Acknowledgements Foremost, I would like to sincerely thank my supervisor, Dr. Leanne Chang, for her great instruction and support in these two years. Her insightful comments and advices on the thesis benefited me to complete this research very much. I am grateful for Dr. Henry Been-Lirn Duh, who provided invaluable opportunities for me to investigate the social impact of AR on collaborative learning in a multi-disciplinary team. His guidance and encouragement along the whole process facilitated me to overcome the challenge in thesis writing. My gratitude also extends to Dr. Vivian Hsueh-Hua Chen who offered constructive suggestions on proposing the research question of the thesis. I am thankful to my partner, Gu Yuanxun. The completion of the thesis was indispensible for his hard work in developing and modifying the system, and great supports for conducting the experiment. Last but not least, I offer my best regards to all the faculty members of CNM. Your kind instructions and academic passion impressed me very much during these two years. i Table of Contents Acknowledgements................................................................................................................ i Table of Contents .................................................................................................................. ii Abstract ............................................................................................................................... iv List of Tables ....................................................................................................................... vi List of Figures..................................................................................................................... vii Chapter 1 Introduction .......................................................................................................... 1 Chapter 2 Literature Review ................................................................................................. 8 2.1 Computer-supported Collaborative Learning (CSCL) ............................................... 8 2.2 Social Interaction in Collaborative Learning........................................................... 12 2.2.1 Theoretical foundations for collaborative learning ........................................ 12 2.2.2 Mutual engagement of social interaction in collaborative learning ................ 17 2.3 Collaborative Learning with Technology-based Scientific Simulation ..................... 25 2.3.1 Mediating functions of technology-based scientific simulation ..................... 25 2.3.2 Potential roles of AR technology in face-to-face CSCL ................................ 29 Chapter 3 Methodology....................................................................................................... 38 3.1 Research Design .................................................................................................... 38 3.2 Participants ............................................................................................................ 40 3.3 Materials................................................................................................................ 41 3.3.1 The systems ................................................................................................. 41 3.3.2 The task ....................................................................................................... 42 3.4 Procedure............................................................................................................... 43 3.5 Data Analysis ......................................................................................................... 44 3.5.1 Quantitative content analysis........................................................................ 45 3.5.2 Conversation analysis .................................................................................. 50 Chapter 4 Results ................................................................................................................ 53 ii 4.1 Quantitative Analyses of the Influences of AR Technology on Social Interaction .... 53 4.1.1 Equality of engagement of social interaction ................................................ 53 4.1.2 Mutuality of engagement of social interaction .............................................. 55 4.2 Qualitative Analyses of the Influences of AR technology on Social Interaction ....... 60 4.2.1 Challenges for mutual engagement of social interaction in collaboration without simulation support ................................................................................... 60 4.2.2 Traditional technology-supported simulations and the characteristics of mutual engagement of social interaction ........................................................................... 70 4.2.3 AR-supported simulation and the characteristics of mutual engagement of social interaction .................................................................................................. 79 Chapter 5 Discussion and Conclusions ................................................................................ 99 5.1 The Influence of AR-supported Simulation on the Equality of Engagement of Social Interaction ................................................................................................................. 100 5.2 The Influence of AR-supported Simulation on the Mutuality of Engagement of Social Interaction ................................................................................................................. 102 5.3 Enhancement of Mutual Engagement of Social Interaction in CSCL..................... 108 5.4 Implications ......................................................................................................... 110 5.4.1 Theoretical implications..............................................................................111 5.4.2 Practical implications................................................................................. 112 5.5 Limitations and Future Research .......................................................................... 113 5.6 Conclusions ......................................................................................................... 114 References ........................................................................................................................ 117 Appendix .......................................................................................................................... 127 iii Abstract As more technologies are integrated with collaborative learning, the mediating functions of technologies on shaping patterns of social interaction in learning activities have received considerable attention in recent years. Mutual engagement of social interaction, being a relational aspect of socially constructing knowledge, is identified as a communication issue to address the efficacy of developing mutual understanding among participants in collaborative learning. Recognized the great potential of AR technology in supporting collaborative learning, this research directs at investigating the influences of an AR-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning for physics. Equality and mutuality of engagement of social interaction serve as two dimensions for measuring mutual engagement of social interaction. 30 pairs of students collaboratively solve the physics problem about elastic collision in one of the three experimental conditions: paper-based, 2D-based or AR-based. The results reveal that the AR-supported simulation does not only possess shared capacities of traditional 2D-supported simulation for promoting the equality and mutuality of engagement of social interaction, but also furthers the enhancement in the mutuality of engagement of social interaction through increasing elaborations and reducing acceptances. Characterized with hybrid attributes of the virtual reality and the real world, the AR-supported simulation enables to motivate collaborators’ mutual engagement in building shared understanding of knowledge by delivering enriched personal experience. This study contributes to the research on the social iv process of CSCL and provides evidence for supporting the promise of AR technology in enhancing face-to-face collaborative learning for physics. v List of Tables Table 3.1 Coding scheme for knowledge-based social interaction ........................................ 49 Table 4.1 Mean (SD) results for equality index .................................................................... 55 Table 4.2 Average percentages of five categories of development statements ....................... 56 vi List of Figures Figure 3.1 The views of the AR-supported simulation, the input interface and the 2D-supported simulation ..................................................................................................... 42 Figure 3.2 The learning scenario of the discussion task ........................................................ 43 Figure 3.3 The scenarios of experiments in the three conditions........................................... 44 vii Chapter 1 Introduction The development of information and communication technologies (ICTs) has great impacts on the whole society. Computer-supported collaborative learning (CSCL), the integration of ICTs with collaborative learning, emerges as a significant field to explore the values of ICTs in fostering learning activities. The social process of collaboration involves participants’ interaction with each other to jointly solve problems; as an integral component of the social process of collaborative learning, social interaction among participants has impacts on the quality of collaborative learning (Roschelle & Teasley, 1995). Nowadays there is an endeavor to explore how social interaction in CSCL can be enhanced by technologies (Kirschner & Kreijins, 2005). Augmented reality (AR) that allows computer-generated virtual objects overlaid onto the physical world has been recognized as the “next generation” pedagogical medium to advance learning quality (Dede, 2008, p.19). With the support of AR technology, multiple learners can not only obtain enriched personal experience brought about by virtual reality through manipulating 3D objects in a shared visual space, but also communicate with each other to solve problems in real-time and real-space. Some researchers assessed the effectiveness of AR applications and found that AR technology entails great capabilities to augment collaborative learning experience (Kaufmann & Dünser, 2007; Klopfer, Perry, Squire, & Jan, 2005; Wagner, Schmalstieg, & Billinghurst, 2006). However, there is still little understanding of the impacts of AR technology on social interaction in collaborative learning. 1 Recognized the potential of AR technology for visualizing scientific phenomena, I plan to explore how AR-supported simulation enhances social interaction in face-to-face collaborative learning for physics in this research. The efficacy of social interaction in collaborative learning is manifested in approaches adopted by collaborators to coordinate the social process for developing shared understanding of knowledge (Erkens, 2004). A relational aspect in building shared understanding, mutual engagement of social interaction, serves as the focus to evaluate the impacts of AR-supported simulation on social interaction in collaborative learning. A collaborative AR system based on mobile phones has been developed to implement interactive simulation of elastic collision for collaboration. The socio-cultural perspective in traditional collaborative learning is incorporated into this study as the theoretical basis to deepen the understanding of the importance of social interaction in CSCL. The socio-cultural perspective emphasizes on the social process in collaborative learning and the communicative function of social interaction in knowledge construction among collaborators; it also highlights the mediating function of surrounding materials and culture in collaborative learning process (Vygotsky, 1978). Individuals’ learning in collaboration is indispensible for active participation in communicating and co-constructing meaning of knowledge among collaborators. Grounded on the socio-cultural perspective, numerous studies have examined the impact of technologies on social interaction in CSCL (Arvaja, 2007b; Suthers & Hundhausen, 2003). However, a majority of them focused on depicting the feature of social interaction in CSCL, but did not assess the enhancement of social interaction supported by technologies (Arvaja, 2007a; Chiu, 2003). Working out a shared solution for a problem serves as a goal of collaborative learning. When 2 jointly constructing knowledge in a task, collaborators have to understand each other along the process of collaboration (Roschelle & Teasley, 1995). Rather than being in a static state, collaborators’ mutual understanding is dynamic and they achieve so during the process of collaborative learning. Indeed, the construction of mutual understanding can be treated as a communication issue (Barron, 2000). The pathway of establishing and maintaining mutual understanding becomes a vital topic to address when studying social interaction in collaborative learning. Although the socio-cultural perspective has contended that tools could mediate the social process of collaborative learning, specifications are needed to identify how tools support social interaction in collaboration. As more technologies are integrated with collaborative learning, understanding the mediating function of technologies in collaborators’ coordination of social process to build mutual understanding is helpful for gaining an insight into the construction of knowledge in CSCL. In order to better understand the effects of AR-supported simulation on strengthening social interaction for developing mutual understanding in collaborative learning, I further apply theory of grounding proposed by Clark and Brennan (1991) to investigate social mechanism underlying the efficacy of social interaction in collaborative learning. Rooted in linguistics, theory of grounding offers an approach to analyze how interpersonal communication takes place to effectively construct mutual understanding (Baker, Hansen, Joiner, & Traum, 1999). Effective interpersonal communication in spoken settings is featured by joint commitments of all participants, which are represented by verbal exchanges that people orient to each other’s statements in the conversation (Clark & Brennan, 1991). The achievement of mutual understanding is not simple accumulations of statements made by 3 participants. In order to ground shared understanding of meaning, after the speaker presents unshared meaning to seek common ground, the addressee needs to build upon it to display his/her understanding. Also, participants need to make joint commitments along the whole conversation to update and develop common grounds. In terms of social interaction in collaborative learning, joint commitments that are reflected by collaborators’ active engagements in initiating new ideas and extending each other’s ideas are important to cultivate mutual understanding of knowledge (Tao, 1999). However, due to ambiguous situations for meaning making of new knowledge, collaborators usually face challenges to make joint commitments when developing shared understanding. Lack of joint commitments in social interaction hinders the construction of mutual understanding and leads to a less ideal solution for solving the group task (Barron, 2003). It is suggested that the level of joint commitments manifested in collaborators’ interaction orientations towards building mutual understanding is a key relational aspect in the social context of collaborative learning, which can affect the effectiveness of working on a shared task. Collaborative relation between participants in the conversation should be a concern when addressing the social process of collaborative learning. Recognized the significance of joint commitments in effective interpersonal communication, extending it to the context of collaborative learning is useful for analyzing the efficacy of social interaction to maintain mutual understanding of knowledge in AR-supported collaborative learning. Mutual engagement of social interaction, emphasizing on joint commitments in social interaction to build mutual understanding of knowledge in peer collaboration, serves as the lens of this research to gain an insight into the impacts of AR-supported simulation on the 4 social process of collaborative learning. Collaboration between peers is one of the important types of collaborative learning in educational practices (Roschelle & Teasley, 1995). A range of research has concentrated on social interaction in dyad to explore the approach of promoting the effectiveness of peer collaboration from different perspectives (Kumpulainen & Kaartinen, 2003; Kumpulainen & Mutanenb, 1999). Mutual engagement of social interaction comprises two dimensions, equality of engagement of social interaction and mutuality of engagement of social interaction (Damon & Phelps, 1989). Equality of engagement refers to the equality between the collaborators to control over the conversation flow through initiating new focuses. Alternatively, mutuality of engagement is the richness of extending each other’s ideas in the course of developing shared understanding (Damon & Phelps, 1989). Effective collaborative learning is featured with both high equality and mutuality of engagement of social interaction. These two dimensions together reveal the mechanism that impacts the formation of mutual understanding of knowledge. The equality of engagement of social interaction exhibits collaborators’ joint commitments to direct the conversation for seeking mutual understanding, while the mutuality of engagement of social interaction is their reciprocal engagements with each other’s ideas to achieve mutual understanding. According to Mercer (1996), social interaction characterized with high mutuality is that collaborators respond to each other’s contributions in a critical but positive way; the level of mutuality of positive acknowledgements and providing supportive information is medium; and simple rejections/no response are low in mutuality. Increasing the use of patterns of social interaction with high mutuality and reducing the use of patterns of social interaction with low mutuality are essential to promote the mutuality of engagement of 5 social interaction. Thus, to assess the effectiveness of AR-supported simulation in enhancing mutual engagement of social interaction, this study attempts to examine the effects of AR-supported simulation on the equality and mutuality of engagement of social interaction in collaborative learning. And patterns of social interaction with different levels of mutuality are identified to facilitate the evaluation of mutuality of engagement of social interaction. In sum, the objective of this research is to investigate the influences of AR-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning for physics. Since equality and mutuality of engagement are two fundamental dimensions of mutual engagement in the social process of collaborative learning, this research examines how AR-supported simulation affects the equality and mutuality of engagement of social interaction in face-to-face collaborative learning for physics. One primary aim of developing new technologies is to address the limitation of traditional technologies and opens up more possibilities to enhance collaborative learning effectiveness. It is necessary to identify new opportunities provided by emerging technologies for supporting the social process of collaborative learning. Hence, in this study, besides examining the capabilities of AR-supported simulation for enhancing mutual engagement of social interaction in face-to-face collaborative learning without simulation support, the comparison between the influences of simulations supported by AR technology and traditional multimedia technology is also conducted to analyze the unique advantage of AR-supported simulation. The contributions of this study are threefold: first, it can broaden the understanding of the mediating role of ICTs in enhancing social interaction of collaborative learning activities; second, it can extend the extant research on social interaction in traditional collaborative 6 learning to CSCL contexts and enrich the approach of analyzing the mechanism underlying the efficacy of social interaction in CSCL; third, it can provide evidence for supporting the value of AR technology in augmenting social interaction in face-to-face collaborative learning. The thesis comprises five chapters. Chapter 2 presents the theoretical background of the research. It begins by providing an overview of the development of research on CSCL. Next, drawing on the theoretical perspective in traditional collaborative learning and theory of grounding in interpersonal communication, it reviews the literature about the importance of mutual engagement of social interaction in collaborative learning. It then discusses previous findings on mediating functions of technology-based simulations in the social process of collaborative learning and the potential role of AR technology in supporting face-to-face collaborative learning for physics. Chapter 3 outlines the methodology of the research. It includes the selection of participants, the materials used, and the procedure of the experiment. Then, the methods of data analyses are introduced. The results of the data analysis are presented in Chapter 4. Chapter 5 discusses the findings of the research. Implications, limitations and future research are also addressed in this chapter. The conclusion of the study is made at the end of the chapter. 7 Chapter 2 Literature Review In this chapter, I start by discussing the research tradition on CSCL, highlighting how ICTs shape social interaction in collaborative learning. To better understand the efficacy of social interaction in CSCL, two theoretical approaches are adopted to study social interaction in collaborative learning and mutual engagement of social interaction in the social process. Next, I review the literature of technology-based simulation on social interaction in collaborative learning. Finally, I introduce the potential role of AR-supported simulations in fostering face-to-face collaborative learning and propose related hypotheses. 2.1 Computer-supported Collaborative Learning (CSCL) ICTs have been increasingly adopted in educational practices with attempts to support collaborative learning. Driven by the advancements of computing technologies in recent years, CSCL, the integration of technologies with collaborative learning, broadens the possibilities of using technologies to complement traditional education in school settings (Zurita & Nussbaum, 2004b). With the advantage of connecting learning contexts and learning activities, CSCL provides a new setting for understanding natures of collaborative learning in pedagogical practices (Kaptelinin, 1999). Despite discussion on technical issues of CSCL, the social influences of this instructional medium have received considerable attention (Fischer & Mandl, 2005; Suthers, 2006). In recent years, there has been a growing effort to investigate the impacts of 8 technologies on the effectiveness of collaborative learning, and their focuses have shifted from learning outcomes to social processes (Chiu, 2003; Stahl, Koschmann, & Suthers, 2006). Regarding CSCL, technologies for supporting collaborative learning are designed to affect the way in which individuals socially construct knowledge and enhance learning effectiveness (Dillenbourg & Fischer, 2007). So far, a great amount of research has concentrated on learning outcomes and investigated the influence of technologies on the effectiveness of collaborative learning by evaluating objective individual learning achievement or group task performance (Reamon & Sheppard, 1997; Sun & Cheng, 2007). Also, there have been some studies using individuals’ perceptions towards learning activities to subjectively measure the effectiveness of CSCL (Alavi, 1994). Rather than attributing knowledge acquisition in collaboration to individual information processing, an increasing number of researchers stressed that the social process of collaborative learning should not be ignored; the social process involves participants’ interpersonal communication with each other to jointly solve problems in collaborative learning (Erkens, 2004; Sangin, Dillenbourg, Rebetez, Bétrancourt, & Molinari, 2008; Stahl et al., 2006). This process-oriented perspective highlights that the efficacy of social interaction is an important aspect to explain the effectiveness of CSCL. “Social affordances”, referring to “properties of a CSCL environment that act as social-contextual facilitators relevant for the learner’s social interaction”, are proposed to address social interaction while building a successful CSCL environment (Kreijns, Kirschner, & Jochems, 2002, p.13). Kreijns et al. (2002) contended that the efficacy of social interaction in CSCL should be emphasized apart from paying attention to the implementation of technology and pedagogy in CSCL; it is crucial to create a CSCL context that motivates 9 collaborators to actively engage in social interaction. The efficacy of social interaction has become a key indicator to assess the success of a CSCL environment (Kirschner & Kreijns, 2005). Thus, examining the social process of CSCL helps to gain an insight into the effectiveness of CSCL. The technologies are not isolated from the social process in CSCL and have the capacity to change patterns of social interaction. Koschmann (2002) proposed that CSCL is “a field of study centrally concerned with meaning and the practices of meaning-making in the context of joint activity and the ways in which these practices are mediated through designed artifacts” (p. 20). The artifacts are important resources for mediating meaning making in CSCL. Individuals in the group can co-construct knowledge by referring to shared learning content provided by technologies (Roschelle & Teasley, 1995). Crook (1998) also placed high value on the resources mediating social interaction in collaborative learning, and pointed out technologies could be significant resources for creating optimal environments by creating referential anchors for collaborators. Linell (1998) proposed the concept of contextual resources to illustrate potential resources that can be used by individuals to negotiate the understanding in social interaction. From the perspective of contextual sources, a range of researchers began studying CSCL as contextual phenomena to explore how social interaction is shaped by technologies, and found that technologies can mediate the process of making sense of problem-solving situations and constructing mutual understanding among participants (Arvaja, 2007a; 2007b). Thus, physical instructional tools are not separated from social interaction in CSCL. It is necessary to explore how technologies influence the efficacy of social interaction in CSCL. 10 Instead of replacing face-to-face communication, a cluster of technologies for supporting co-located collaborative learning has been developed for enhancing the efficacy of face-to-face interaction in the learning process (Reamon & Sheppard, 1997; Zurita & Nussbaum, 2004b). The research on the impact of incorporating computing technologies into face-to-face collaborative learning provided supportive evidence that computing media can create innovative environments for augmenting social interaction in real-time and real-space (Zurita & Nussbaum, 2004a; 2004b). Despite the widespread use of networked technologies to support distributed collaborative learning, face-to-face collaboration among peers is popularly used in educational settings. It has gained rising attentions as more ICTs are introduced to school environments (Liu, Chung, Chen, & Liu, 2009). Thus, face-to-face collaborative learning serves as an important context to examine how technologies affect social interaction in CSCL. Although technologies show potential for supporting social interaction in collaborative learning, integrating technologies with collaboration does not guarantee desired outcomes. Evaluating the effectiveness of CSCL is needed. Media characteristics are capable of affecting learning practices (Lai, Yang, Chen, Ho, & Chan, 2007). Thus, media characteristics need to be taken into account when applying technologies to collaborative learning. As more emerging technologies are developed for promoting collaborative learning, assessing the influence of technologies on collaborative learning is critical to further exploit the capability of technologies for enhancing learning activities. Meanwhile, considerations should be given to the method of measuring the efficacy of social interaction in CSCL. There are growing interests on the impact of technologies on the social process of CSCL, however, a 11 large portion of them only proposed instruments to characterize types of social interaction in a single CSCL context or across different contexts (Chiu, 2003; Sangin et al., 2008). Although this approach facilitates to understand how technologies shape patterns of social interaction in CSCL, it fails to evaluate how technologies influence the quality of social interaction and thus provides limited knowledge on the effectiveness of technologies to foster social interaction in collaborative learning. Adopting the features of social interaction that could reflect the quality of social interaction is significant for analyzing the impact of technologies on the efficacy of social interaction in collaborative learning. Since the research tradition on CSCL is relatively new, there is a need to integrate relevant works conducted within traditional collaborative learning and along with those on interpersonal communication contexts in order to better understand the feature of social interaction in face-to-face CSCL. 2.2 Social Interaction in Collaborative Learning This section has two parts. First I present the theoretical approach to explain the role of social interaction in collaborative learning and the underlying mechanism that influences the efficacy of social interaction. Next, I describe the significance of treating mutual understanding as a communication problem in collaborative learning. Then I proceed to a review of mutual engagement of social interaction and discuss the rationale of adopting it to evaluate the efficacy of social interaction in collaborative learning. 2.2.1 Theoretical foundations for collaborative learning Studying social interaction in collaborative learning has become a key issue in the research agenda of collaborative learning nowadays. Traditionally, individual functioning was 12 stressed while social interaction among participants was identified as an external environment for individuals to acquire knowledge in collaborative learning (Dillenbourg, Baker, Blaye, & O’Malley, 1996). Since individuals are integral parts of a group and they need to interact with each other to jointly solve problems, ignoring social interaction limits the understanding of group functioning in collaborative learning. In recent years, the salience of social process in collaborative learning is much more emphasized (Erkens, 2004). An increasing amount of literatures have studied the social process in collaboration through identifying characteristics of social interaction in different conditions or the effects of types of social interaction on learning outcomes (Barron, 2003; Hogan, Nastasi, & Pressley, 1999). The socio-cultural perspective serves as a theoretical basis for interpreting the social process of constructing knowledge and the mediating role of external circumstances in collaborative learning (Vygotsky, 1978). This perspective posits that the meaning of knowledge is built on shared speech, tools and activities. It explains the significant role of social interaction in cognitive growth by stressing participation in knowledge construction among collaborators in learning activities. The communicative functions of social interaction are highlighted rather than simply considering its functional role as a catalyst in fostering mental development (Barron, 2000). Additionally, this perspective explains the mediating function of material tools and culture in affecting individuals’ social interaction to jointly construct knowledge in collaboration. Generally, the socio-cultural approach suggests that integrating social and material surroundings with collaborative learning is significant to understand how knowledge construction is socially shaped in problem-solving (Arvaja, 2007a). 13 The socio-cultural perspective offers a foundation for analyzing the critical role of social process of constructing knowledge in collaborative learning. It has inspired a body of studies to explore social interaction in collaborative learning by focusing on different aspects of the social process (Barron, 2003; Staarman, Laat, & Meijden, 2002). A majority of them investigated the meaning of individual statements rather than incorporate the meaning linkage between individual statements into the analyses (Chin & Brown, 2000; Russell, Lucas, & McRobbie, 2004). They typically examined the content of individual statement based on the depth of cognitive processing of knowledge with attempts to gain an understanding of how collaborators construct knowledge (Chin & Brown, 2000). Whereas assessing the quality of each individual statement is helpful to evaluate the cognitive approach used by individuals in collaborative learning, examining social interaction at the group level by taking interaction sequences into account could give insight into the process of meaning negotiation. In recent years, more attention has shifted to the establishment of mutual understanding in collaborative learning (Barron, 2000; Erkens, 2004). Developing mutual understanding among participants is acknowledged as the heart of collaborative activities, and the dynamic nature of mutual understanding provides a basis for further addressing relevant social interaction issues. Roschelle and Teasley (1995) defined collaboration as “a coordinated, synchronous activity that is the result of a continued attempt to construct and maintain a shared conception of a problem” to heighten the dynamics of mutual understanding and its significant role in collaborative learning (p. 70). It is suggested that the pathway of achieving mutual understanding is a key aspect to assess the effectiveness of collaborative learning. Solving a problem towards a shared goal among participants is an 14 aim of collaborative learning, and thus participants need to reach mutual understanding of solutions for tackling the problem. The efficacy of social interaction in collaborative learning is manifested in individuals’ communicative strategies to coordinate social interaction for developing mutual understanding. The moment-by-moment feature of social interaction has been identified by a growing number of researchers (Erkens, 2004; Kumpulainen & Mutanen, 1999). They contended that the continuous interpretive process among collaborators is needed to construct mutual understanding of knowledge. Since constructing mutual understanding is a social dynamic activity along the collaboration, it is necessary to concentrate on social interaction of collaborative learning and explore how collaborators manage to build mutual understanding of knowledge in the social process. While the socio-cultural perspective identifies the significance of social interaction in constructing knowledge in collaborative learning, theory of grounding, explicitly addressing the association of joint commitments in interpersonal communication with the development of mutual understanding, is useful for furthering the understanding of social mechanisms underlying the achievement of mutual understanding in collaborative learning (Baker et al., 1999). Rooted in linguistic research, grounding refers to the interpersonal communication process of developing shared understanding among participants through verbal interaction (Clark & Brennan, 1991). It is suggested that communicative acts, the coordination of participants’ utterances, function as media for people to negotiate their understanding in conversation. Based on the viewpoint of Clark (1996), communicative acts are inherent joint actions in collaborative activities, which result in accumulations of shared understanding through coordinating the social process. Joint commitments of all participants are required to 15 effectively build shared understanding in the conversation (Clark, 1996). And the construction and maintenance of common grounds are indispensible for joint efforts of individuals involved in. Regarding collaborative learning, verbal interaction is perceived as “a social mode of thinking”, which serves to be a medium of jointly constructing knowledge rather than merely sharing one’s own thoughts with each other (Mercer, 1996, p. 374). Not only do collaborators verbally exchange their own understanding on the problem, but also evaluate and reflect on others’ ideas and negotiate meaning in the conversation. According to the perspective of grounding, meaning negotiation is indispensible for social interaction engaged by participants, and joint commitments should be stressed when studying collaborators’ social interaction (Clark, 1996; Clark & Brennan, 1991). Establishing and maintaining mutual understanding of knowledge are significant to the effectiveness of collaboration. Thus, extending the investigation of joint commitments in social interaction to the context of collaborative learning is beneficial for attaining an insight into the efficacy of the social process to construct mutual understanding of knowledge in collaborative learning. There are series of challenges for participants to jointly construct mutual understanding in collaborative learning, and collaborative relation represented by participants’ interaction orientations to build mutual understanding relates to the effectiveness of collaboration. When students are required to solve a problem on new knowledge together, they have to cope with some challenges in the process of reaching shared understanding, for instance, collectively carrying out explorations, interpreting ambiguous situations, negotiating socio-cognitive conflicts and refining shared cognition (Barron, 2000; Roschelle, 1992). Storch (2002) found that not all groups in collaborative activities behave in a collaborative 16 manner, and low level of joint commitments have impacts on the effectiveness of achieving mutual consensus in meaning negotiation. Based on the view of Barron (2003), joint commitments reflected by collaborators’ interaction orientations to each other in the group should not be ignored since this relational aspect can significantly affect the success of developing mutual understanding of knowledge in collaborative learning; identifying opportunities and challenges for motivating joint efforts in social interaction of collaborative learning is important to foster the effectiveness to build mutual understanding. Therefore, it is necessary to examine social interaction in collaborative learning through the lens of joint commitments to better understand the characteristics of the social process in collaborative learning as well as the opportunities and challenges for developing mutual understanding. Mutual engagement of social interaction, stressing joint commitments in social interaction for reaching mutual understanding, has emerged as a vital relational issue in the research on social interaction of collaborative learning (Damon & Phelps, 1989). In the following part, relevant literatures on mutual engagement of social interaction in collaborative learning are reviewed. 2.2.2 Mutual engagement of social interaction in collaborative learning With respect to peer learning, equality and mutuality of engagement of social interaction are two fundamental dimensions used to illustrate mutual engagement of social interaction (Damon & Phelps, 1989). Peer learning is a significant way to motivate students to actively construct knowledge through negotiating the understanding of tasks with each other within a small group (Kumpulainen & Kaartinen, 2000). A body of studies has focused on 17 characterizing social interaction between peers in order to improve the effectiveness of peer learning (Kumpulainen & Kaartinen, 2003; Roschelle & Teasley, 1995). Equality and mutuality of engagement of social interaction are two vital aspects to examine social interaction in peer learning (Damon & Phelps, 1989). Equality refers to the level of controlling over the group task, which is represented by the control of the direction of social interaction between two people (Damon & Phelps, 1989). Hence high equality indicates that both parties in a group actively engage in controlling over the flow of social interaction rather than only one party dominates the conversation. Mutuality is described as the degree of engagement with the contribution of the partner’s (Damon & Phelps, 1989). In terms of high mutuality, it is characterized with rich extension of each other’s statement between two parties in a group. Achieving high equality and high mutuality of engagement of social interaction are the objective of constructing effective collaborative learning environments. Interaction sequences in conversation serve as an important context to analyze the equality and mutuality of engagement of social interaction in collaborative learning. Barnes and Todd (1977) distinguished four features of the social process in collaborative learning, which includes initiating new focus of topic, eliciting information from others, building upon preceding ideas, and qualifying the disagreement and complexity in previous utterances. They proposed that the concept of collaborativeness, referring to “links between succeeding utterances”, to illustrate joint actions engaged by individuals within a small group (p. 3). Initiating a new focus functions as a shared frame in the conversation, and extending and qualifying statements are needed in order to sustain the development of social interaction in collaborative learning. Hogan et al. (1999) proposed “interaction sequence” to define the flow 18 of social interaction, which is “a series of turns bounded by statements that initiate a new level of focus” (p.388). The formation of an interaction sequence is constituted with two elements, initiating a new level of focus in the conversation and extending this focus at the same level. After one person presents a statement at a new level of focus, the response to the initiation at the same level of focus from other group members is required to build an interaction sequence. Then, the conversation centering upon this level of focus contributes to the richness of interaction sequence. And mutual understanding is developed during the extension of the initiating statement. An interaction sequence ends when one group member changes the previous level of focus to a different one. In the context of an interaction sequence, the statement that initiates of a new level of focus plays the role of a controller of conversation direction, which is essential to measure the equality of engagement of social interaction, while the statements that extend this level of focus could be used to characterize the mutuality of engagement of social interaction (Galaczi, 2004). So, analyzing the initiating statement and the development statements in an interaction sequence facilitate to get a deep insight into mutual engagement of social interaction in collaborative learning. The equality of engagement of social interaction highlight the relationship of peers formed in the social process of collaborative learning for seeking mutual understanding (Damon & Phelps, 1989). Different levels of equality are evaluated by collaborators’ equality of initiating a statement with a new focus in the conversation (Galaczi, 2004). The members in dyads with high equality play relatively equal roles in opening new focuses in the process of collaborative learning. However, high equality of engagement of social interaction does not guarantee group functioning in a collaborative manner. Galaczi (2003) found that dyads might 19 behave in a “solo vs. solo” pattern, which indicates that both participants actively introduce new focuses but do not like reacting to each other’s ideas and further expanding them in the conversation (p.2). This type of dyads has a high level of equality of engagement of social interaction, while the level of mutuality of engagement of social interaction is low. Hence, apart from high equality, effective collaborative learning should be characterized with high mutuality at the same time. The behavior of opening a new focus is only the beginning for establishing mutual understanding and responses to it is needed to reach mutual understanding (Clark & Brennan, 1991). Research usually links communicative functions of social interaction to analyze the mutuality in the process of constructing mutual understanding in collaborative learning. Two basic patterns of social interaction are identified to heighten the importance of mutuality when forming shared understanding in collaborative learning, “construction and co-construction of meaning”, and “constructive conflict” (Bossche, Gijselaers, Segers, & Kirschner, 2006, p.495). The introduction of specific meaning of a situation or an approach to solve the problem is described as construction of meaning, and this will generate co-construction of meaning among group members, resulting in new understanding of the situation within the group. The construction and co-construction of meaning is not merely the aggregation of independent meaning inserted by individuals, but the integration of meaning and the achievement of mutual understanding based on proceeding negotiations. Also, construction and co-construction of meaning is not the only path of building mutual consensus (Fischer, Bruhn, Gräsel, & Mandl, 2002). When the viewpoints expressed by individuals are different, further negotiations on the meaning are needed to obtain mutual agreement. Constructive conflict is 20 defined as tackling different views arisen from the conversation with attempts to reach reciprocal understanding (Fischer et al., 2002). Empirical evidence has also been offered to support the vital role of construction and co-construction of meaning, and constructive conflicts in developing mutual understanding of knowledge (Bossche et al., 2006). Identifying construction and co-construction of meaning and constructive conflicts in collaborative learning can yield insights into general mechanisms underlying the mutuality of engagement of social interaction. However, a body of research only relied on construction and co-construction of meaning or constructive conflicts to assess mutuality without taking the variability of mutuality within co-construction of meaning and constructive conflicts into consideration (Barron, 2003; Fischer et al., 2002). Since social interaction patterns reveal different levels of mutuality, a broad conceptualization of mutuality makes it difficult to investigate detailed characteristics of social interaction on the path of achieving consensus. So it is crucial to identify patterns of social interaction with different levels of mutuality in collaborative learning. Indeed, the central issue of enhancing mutuality of engagement of social interaction is increasing the use of patterns of social interaction with high mutuality and reducing the use of patterns of social interaction with low mutuality. Interaction patterns featured by different levels of mutuality have impacts on the effectiveness of constructing mutual understanding in collaborative learning. Mercer (1996) distinguished three major interaction patterns for analyzing the quality of social interaction in collaborative learning based on communicative functions, which includes “exploratory talk”, “cumulative talk” and “disputational talk” (p.369). They represent different ways that collaborators build on each other’s statements in the social process. Exploratory talk occurs 21 “when partners engage critically but constructively with each other’s ideas” (p.369), such as presenting arguments, proposing alternative hypothesis and asking for clarifications. Cumulative talk refers to “speakers build positively but uncritically on what the other has said” (p.369), which is manifested in confirmations and repetitions. Disputational talk usually takes the forms of “disagreement and individualized decision making” (p.369). Among these types of talk, exploratory talk is conceived as the most productive one in group learning activities that strengthens the reciprocity in social interaction and thereby promotes shared knowledge construction (Mercer, 1996). By incorporating collaborators’ reciprocal engagements into the analysis of social interaction, Mercer’s (1996) categories lay a foundation for identifying social interaction patterns with different levels of mutuality in the social process of collaborative learning. The integration of joint commitments in social interaction with shared knowledge construction reveals the social process underlying the development of mutual understanding of knowledge in collaboration. On the basis of Mercer’s (1996) instrument, some researchers sought to further operationalize social interaction in order to examine the features of mutuality of engagement of social interaction in collaborative learning (Barron, 2000; Visschers-Pleijers, Dolmans, Wolfhagen, & Van der Vleuten, 2005). Among these, Barron (2000) developed five main types of responses to new proposals in the social process of collaborative learning, namely, “acceptances”, “clarifications”, “elaborations”, “rejections” and “no response”, to investigate how collaborators coordinate verbal interaction for establishing and maintaining mutual understanding (p.414). For these types of responses, elaborations, characterized with offering extra information, advices and justifications, have a high level of mutuality. Clarifications are 22 proposing follow-up questions. They are also high in mutuality since the person has to integrate new meaning of the knowledge stated by the partner with his/her prior understanding to make clarifications. Acceptances include simple agreements or repeating prior statement. Rejections and no response belong to “disputational talk” mentioned above, which represent a low level of mutuality and exert negative influences on mutual understanding development (Mercer, 1996, p.369). After analyzing communicative functions of social interaction in collaborative learning based on this scheme, Barron (2000) found that the variability of levels of mutuality exists across different groups, which affects the efficacy of building understanding among collaborators. In particular, elaborations do contribute to effective coordination among group members to reach mutual understanding, while rejections and no response hinder the construction of mutual understanding. This scheme gives an insight into interaction patterns with different levels of mutuality of engagement of social interaction in collaborative learning. However, the limitation of this study is that it did not separate solely adding positive information from those critical responses. Accumulating information benefits building common grounds, but the contribution to the other’s statement is limited and the reciprocity is lower than critical elaborations. Since exploratory talk is highly appreciated to enhance the quality of social interaction, it is necessary to examine critical elaborations and accumulations respectively. Also, although acceptances have positive effects on quickly reaching consensus among collaborators, they are less constructive to develop an understanding at a higher level since little new information about the meaning of knowledge is added (Gijlers, Saab, Van Joolingen, De Jong, & Van Hout-Wolters, 2009). Fostering the richness of accumulative talk is important to promote the mutuality of 23 engagement of social interaction. Therefore, increasing critical elaborations, clarifications, accumulations and reducing simple acceptances, rejections and no responses are significant to enhance the mutuality of engagement of social interaction in collaborative learning. Extending mutual engagement of social interaction to the research on CSCL is helpful for enriching the understanding of the influence of technologies on the efficacy of social interaction in CSCL. Social interaction characterized with a high level of mutual engagement relates to the effectiveness of reaching mutual understanding of knowledge in collaborative learning. Nowadays, working in pairs is a commonly used form of collaborative learning in school environments. Since the levels of equality and mutuality of engagement of social interaction are important to create a constructive collaborative learning context, it is necessary to explore the impact of technologies on mutual engagement of social interaction in face-to-face collaborative learning. Tao (1999) has introduced mutual engagement of social interaction into the research of CSCL for science subjects. He used equality (high/low) and mutuality (high/low) to qualitatively evaluate the engagement of each dyad. But simply utilizing a continuum with a range of low to high level is inadequate to identify the salient features of social interaction, making it difficult to understand the mechanism underlying the construction of mutual understanding. Hence, more considerations should be given to the approach of characterizing and evaluating equality and mutuality of social interaction in CSCL. On the basis of literatures on collaborative learning, interaction sequences are regarded as vital contexts to examine the equality and mutuality of engagement of social interaction (Galaczi, 2004; Hogan et al., 1999). Initiating statements and development statements in interaction sequences are examined in this study to capture the features of 24 equality and mutuality of engagement of social interaction. Also, to better understand the efficacy of technologies, the patterns used to analyze mutuality of engagement of social interaction should be able to indicate social interaction with different degrees of mutuality. Therefore, the instrument developed by Barron (2000) is applied and modified for this study. 2.3 Collaborative Learning with Technology-based Scientific Simulation This section is organized into two parts. I first present the benefits of technology-based simulations for science subjects and review the literature on the impacts of collaborative use of simulations on mutual engagement of social interaction. Then I proceed to introduce the promising role of AR-supported simulation for enhancing face-to-face collaborative learning. 2.3.1 Mediating functions of technology-based scientific simulation The simulation is a typical genre of computing technologies applied in face-to-face CSCL. “Learning with simulations” is conceived as one of the important impacts of technologies on science education, which is able to provide more pedagogical opportunities for students to actively explore and acquire knowledge (Webb, 2008, p. 134). In order to make better use of technology-based simulations, it is suggested that more research is needed to investigate to what extent such kinds of applications benefit learning activities. Simulations reveal great potential for supporting learning activities of science subjects. Technology-based simulations are largely used to convey meaning in abstract and complex education practices through visualizations, which contribute to making sense of the knowledge and enhancing the quality of peers’ interaction (Reamon & Sheppard, 1997; 25 Roschelle & Teasley, 1995). Regarding science subjects full of abstract information, interactive visualizations of science phenomena make it more possible for learners not only to access concrete information to comprehend conceptual knowledge, but also to provide an exploratory tool to construct knowledge based on hands-on experiences (Shaer, Kol, Strait, Fan, Grevet, & Elfenbein, 2010). Especially at the initial stage of learning science knowledge that needs high-order information processing, the opportunity of conducting experiments are invaluable to understand scientific phenomena and principles that cannot be directly observed in the real world (Järvelä, Bonk, Lehtinen, & Lehti, 1999). The collaborative use of technology-based simulations in learning activities attracts more attentions from designers and researchers these years as the value of CSCL are acknowledged by an increasing amount of literatures (Chee & Hooi, 2002; Sangin et al., 2008). Great importance has been attached to the shared experience in collaborative learning (Pauchet et al., 2007). It is suggested that the shared experience of manipulating artifacts can broaden common grounds among collaborators, which fosters to build mutual understanding in problem solving by referring shared artifacts. In recent years, experimenting with visual simulations has been widely used to support collaboration in science education with an attempt to motivate students to propose scientific inquiries and stimulate learning interests. There are some studies on the benefits of technology-based simulations in science domain focusing the attention on the influence of simulations on individuals’ conceptual understanding of knowledge in collaborative activities (Reamon & Sheppard, 1997; Whitelock et al., 1993). As the significance of social interaction in collaborative learning is increasingly acknowledged, the effects of technology-based simulations on social interaction 26 in collaborative science learning become crucial to the research on CSCL (Colella, 2000; Tao, 1999). Technology-based simulations reveal a promising role in mediating social interaction to build mutual understanding of meaning in collaborative learning. Within the setting of collaborative science learning, shared visual information displayed by the simulation can be significant resources for social interaction instead of merely providing external representations of knowledge to assist the demonstration of scientific concepts and principles (Andrews, Woodruff, MacKinnon, & Yoon, 2003). A number of researchers interpreted the role of visual information in mediating social interaction and establishing mutual understanding of meaning in collaborative learning (Rochelle & Teasley, 1995; Suthers, 2006). For example, Rochelle and Teasley (1995) analyzed the mediating function of a scientific simulation in the process of collaboratively solving a physical problem and concluded that experimentations serve as a means to coordinate meaning negotiation by continuously providing shared resources and references. They claimed that shared activities supported by the interactive simulation encourage collaborators generating new ideas, refining the prior understanding and resolving conflicts, which in turn benefit the accumulation of shared understanding of scientific knowledge. Järveläet al. (1999) also noted that the simulation is a shared referential anchor for individuals in small groups to negotiate the meaning of new knowledge in a more reciprocal manner. They contended that it is significant to provide interactive instructional supports to knowledge learning since reciprocal understanding during interpersonal communication have facilitation effects on learning effectiveness. Furthermore, embedding scientific simulations supported by technologies in the 27 context of collaborative learning is regarded as a method to stimulate collaborators’ mutual engagement of social interaction to form shared understanding (Järvelä et al., 1999). “Representational guidance” is used to illustrate the role of external representations in shaping social interaction in CSCL, which exerts positive influences on expressing, explaining and refining ideas in collaboration due to “ease of reference” and “reminding” (Suthers, 2001, p. 260). Suthers (2001) claimed that collaborators prefer to elaborate on the knowledge that is salient in their shared context, and emerging knowledge yielded from common experiences of manipulating external representations is useful for organizing the conversation in collaborative learning. Suthers and Hundhausen (2003) expanded this point of view and systematically clarified three main functions of external representations based on the process of jointly negotiating the meaning in collaborative learning, which provided a deep insight into the mechanism underlying the effects of external representations on mutual engagement of social interaction. The three functions include initiating the negotiation of meaning, creating a shared reference in mean-making of the situation, and offering group memory for furthering elaborations. Interacting with external representations can motivate people to develop new ideas about the topic. When individuals in the group have a new idea about shared representations, they feel obliged to initiate the conversation and negotiate meaning with others to seek mutual understanding of knowledge. During the process of meaning negotiation, the shared experience of using external representations and observing subsequent effects of the manipulation can engage collaborators in reflecting on the information presented by external representations and discussing related topics. Additionally, external representations function as shared memory of collaboration since collaborators can 28 refer to prior information offered by external representations, which afford them to modify their solutions to the problem over time. Therefore, external representations open up new possibilities for encouraging both initiations of topics and extensions of meaning in ongoing conversation to construct mutual understanding. As a kind of external representation of scientific phenomena, technology-based simulations enable to offer shared referential resources to motivate collaborators to initiate new focuses related to the phenomena and further negotiate meaning of it, which are key aspects of mutual engagement of social interaction in collaborative learning. Even though technology-based simulations generally show positive impacts on the process of collaborative learning, simulations might be ineffective to improve learning quality. The mediating role of simulations characterized by different visual representations in social interaction of CSCL can be various (Reamon & Sheppard, 1997). The capabilities of technology-based simulations for supporting mutual engagement of social interaction should be clearly identified to gain an insight into the opportunities that contribute to the enhancement of social interaction in collaborative learning. As more emerging technologies are developed for implementing scientific simulations in collaborative learning, it becomes important to concentrate on analyzing the influence of collaborative use of these new applications on mutual engagement of social interaction in learning practices. 2.3.2 Potential roles of AR technology in face-to-face CSCL Continuing advancements of technologies afford new possibilities for supporting face-to-face CSCL in small groups. This type of emerging technologies directs at augmenting 29 face-to-face interaction among co-located collaborators in the learning process by addressing the limitation of traditional technologies. In recent years, more new interactive simulation tools supported by emerging technologies have been developed to introduce innovative experience of collaborative science learning (Cole & Stanton, 2003; Colella, 2000). AR is an emerging interactive medium whereby virtual graphics overlay physical objects in the real world in real time. More recently, an increasing amount of explorations have been carried out to apply AR technology for supporting collaborative learning activities, and AR is recognized as a powerful tool to facilitate face-to-face collaborative learning (Kaufmann, Schmalstieg, & Wagner, 2000). Building on the characteristics of traditional multimedia technology, AR shares the capacity of implementing interactive simulations of scientific phenomena. It can simulate scientific phenomena interactively and allow users to constructively explore abstract knowledge of science subjects by running simulations. Supported by collaborative AR technology, multiple users are able to manipulate three-dimensional objects to do collaborative tasks while interacting with each other in a face-to-face setting. To date, a range of AR-supported scientific simulations have been developed to aid the education of science subjects such as math, physics, and chemistry (Kaufmann et al., 2000; Weghorst, 2003). Meanwhile, the unique interface of AR possesses the capability to construct a more engaging collaborative learning environment compared to traditional multimedia technologies. Entailing the power of bridging real and virtual environments, AR-supported collaborative learning environments have a mixture of attributes of the virtual reality and the real world, which facilitate to create enriched hybrid learning experience for students in collaboration. 30 On the one hand, AR reveals great potential for strengthening personal experience of collaborators’ by letting them deeply involved in scientific simulations. One of the most significant strengths of virtual reality is providing a more situated learning context that enables individuals to gain first-person experience in learning activities (Avradinis, Vosinakis, & Panayiotopoulos, 2000). Instead of being external agents, users embed themselves in virtual learning environments and engage in knowledge building based on their personal and direct experience instead of the description of knowledge delivered by a third party (Winn, 1993). In the context of AR, despite it does not create a fully immersive virtual environment, virtual objects superimposed on the physical world are still able to exert some similar effects as virtual reality on strengthening personal experience in learning activities (Shelton & Hedley, 2004). Enriched personal experience contributes to individuals’ involvement in learning scenarios, and stimulates their interests and motivations for deep learning (Colella, 2000). As noted above, technology-based simulations function as referential resources for encouraging students exploring science phenomena and elaborating on the knowledge based on hands-on experience. However, one limitation of most technology-based simulations is the boundary between the personal experience and the simulation, which makes it difficult for users to take a first-person perspective towards the simulation in the exploratory learning process (Avradinis et al., 2000). People tend to feel that they are only audiences of the simulation rather than an integral part of the simulation scenario. The subjective linkage between users and simulated scenarios is weak, which may reduce their involvements in the learning practices. In the context of collaborative learning, individuals’ subjective tie with the simulated setting can affect their engagement of exploring scientific knowledge (Colella, 31 2000). Incorporating virtual reality into learning practices provides a well-suited context to address this constraint of traditional multimedia technology and strengthen learners’ subjective participatory sense in learning scenarios. Recognized the salient role of the sense of subjectivity in constructing mutual understanding in collaborative learning, Suthers (2006) suggested that the mediating role of technologies in reflecting subjectivity in the learning process should be considered in future design. He also proposed the term of “reflector of subjectivity” to describe the opportunities offered by technologies to foster the development of mutual understanding of knowledge in collaborative learning (p.328). On the other hand, compared to collaboration in fully immersive virtual reality where individuals participate in communication in a virtual-mediated manner, collaboration supported by AR technology allows people to synchronously interact with each other in real world, which has positive effects on exchanging ideas and reflecting on simulations. The face-to-face communication channel makes it easy and convenient for collaborators to elaborate on prior simulations and develop deeper understanding of the scientific principle underlying the simulations. Colella (2000) clarified that the primary benefit of strengthened personal experience in collaboration is not to let collaborators attain immersive learning experience, but to motivate them to constructively analyze the situation and the underlying principles at the stage followed with the immersive experience. Hence, collaborators’ social interaction in face-to-face situation contributes to the realization of facilitation effects of strong subjective tie between personal experience and learning scenarios in enhancing their engagement to explore scientific knowledge. Combining the advantages of virtual reality in fostering personal experience and the 32 significant role of face-to-face interaction in providing natural means for elaborating on the situation in collaborative learning, AR demonstrates as a promising interface to promote the subjective tie between learning experience and simulation scenarios, and transform deep personal involvement to high engagement of social interaction. So, it is plausible to assume that AR-supported simulations can provide unique opportunities for supporting mutual engagement of social interaction in collaborative learning compared with traditional multimedia technologies. Specifically, AR-supported simulations show great potential for increasing the use of patterns of social interaction high in mutuality and reducing the use of patterns of social interaction low in mutuality. Evaluating the effectiveness of AR technology in collaboration has become an integral area in AR research (Billinghurst, 2008). To better understand the value of AR technology for supporting collaborative learning, more recent research were dedicated to measuring the effectiveness of collaborative AR applications and the findings indicated that AR technology has positive effects on learning experiences (Cole & Stanton, 2003; Klopfer, Perry, Squire, & Jan, 2005). However, most of the studies relied on learning outcomes or subjective measurements of learning experience to assess the effectiveness of AR-supported collaboration (Costabile, De Angeli, Lanzilotti, Ardito, Buono, Pederson, 2008; Wagner, Schmalstieg, & Billinghurst, 2006). Very few studies have evaluated the mediating effects of AR technology on enhancing social interaction in collaborative learning. The understanding about the influence of AR technology on the enhancement of social interaction in collaborative learning is limited. For example, Klopfer et al. (2005) used descriptive qualitative analysis to examine social interaction in a co-located role play educational game 33 supported by AR technology. They found that the application could encourage users sharing information with each other and actively solving the problem together. However, no comparison was made between AR and other interfaces to indicate the superiority of AR technology in supporting social interaction in collaborative learning. Besides incorporating basic features of traditional technologies to facilitate certain activities, emerging technologies should provide additional features to overcome the constraints of traditional technology and enhance the effectiveness of technology to support the activities. AR serves as a relatively new medium in the field of CSCL. Identifying the unique advantages of AR interface compared with traditional multimedia interfaces is significant to further promote the usage of AR in collaborative learning. Since the effectiveness of technologies in supporting social interaction is important to build a successful CSCL environment, it is necessary to examine how AR affects social interaction in collaborative learning. Thus, AR technology demonstrates great potential for extending collaborative learning experience for science subject. It is crucial to examine the influences of AR-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning to gain an insight into how AR technology enhances the efficacy of social interaction in face-to-face collaborative learning without simulation support. Augmenting the capability of traditional technologies is a driver for developing new technologies. Besides possessing the benefits of simulation supported by traditional multimedia technologies in collaborative learning, the simulation supported by emerging AR technology should have more advantages for fostering the effectiveness of learning activities. Thus, it is necessary to simultaneously analyze new opportunities offered by AR technology 34 to strengthen mutual engagement of social interaction compared with traditional multimedia technologies. Based on the review above, two research questions were proposed: RQ1. What influences does AR-supported simulation have on the equality of engagement of social interaction in face-to-face collaborative learning for physics? RQ2. What influences does AR-supported simulation have on the mutuality of engagement of social interaction in face-to-face collaborative learning for physics? In this study, six hypotheses are developed to assist to answer the research questions. All hypotheses are tested based on three conditions of face-to-face collaborative learning for physics, including conditions without simulation tool, with traditional multimedia technology-supported simulation and with AR-supported simulation. For the influence of AR-supported simulation on the equality of engagement of social interaction raised in RQ1, I predicted that: H1: There are significant differences in the equality of engagement of social interaction in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported simulation. To investigate the influences of AR-supported simulation on the mutuality of engagement of social interaction proposed in RQ2, five hypotheses are developed separately. Elaborations, clarifications, accumulations, acceptances, and rejections/no response are five typical patterns of social interaction featured with different levels of mutuality in the social process of collaborative learning. The use of each pattern of social interaction represents the 35 approach adopted by collaborators to construct mutual understanding of knowledge in collaborative learning. The following hypotheses are proposed to examine the impacts of AR-supported simulation on patterns of social interaction with different levels of mutuality. H2a: There are significant differences in using elaborations in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported simulation. H2b: There are significant differences in using clarifications in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported simulation. H2c: There are significant differences in using accumulations in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported simulation. H2d: There are significant differences in using acceptances in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported simulation. H2e: There are significant differences in using rejections/no response in face-to-face collaborative learning for physics across three conditions: (1) without simulation support; (2) with traditional multimedia technology-supported simulation; (3) with AR-supported 36 simulation. 37 Chapter 3 Methodology In this section, I start by presenting the overall research design of this research. Next, the background information of participants is described. The materials adopted in this research, including the system and the task, are introduced. Then, I outline the procedure of conducting experiments and the approach of analyzing data. 3.1 Research Design This research seeks to investigate the influences of an AR-supported simulation on two primary dimensions of mutual engagement of social interaction in face-to-face collaborative learning for physics, equality and mutuality of engagement of social interaction. The experimental design dominating the research tradition of CSCL was adopted in this research (Stahl et al., 2006). The single-factor between-subjects design with three conditions featured by different instructional media was used. Elastic collision, an important phenomenon in the instruction of conservation of momentum in physics of Junior College in Singapore, was chosen as the learning scenario of the collaboration. The collaboration was in the form of face-to-face dyadic discussion. The first condition was normal face-to-face collaborative learning without simulation support, and pairs of students in this condition discussed the question with the support of paper-based instructional material (“Paper-based condition”). In the second condition, a simulation system supported by 2D graphics technology was provided to facilitate the group discussion (“2D-based condition”). In the 38 third condition, a simulation system supported by AR technology was offered to aid the discussion (“AR-based condition”). All participants were divided into two-member groups and then each group was randomly assigned to one of the three conditions. The requirement for being a participant was that he/she had no prior knowledge of elastic collision. Open-ended questions were adopted in the discussion since they could provide more opportunities for motivating collaborators to explore the knowledge. Three main considerations were given for the experiment design. First, the capability of AR-supported simulation for fostering mutual engagement of social interaction in normal face-to-face collaborative learning for physics should be examined. Promoting the effectiveness of collaborative learning without simulation support is a basic requirement for a simulation tool to meet, so the comparison between the collaboration without simulation support and with the AR-supported simulation was made. Second, it is necessary to analyze the advantages of AR-supported simulation for supporting mutual engagement of social interaction in face-to-face collaborative learning compared to those of traditional technologies. Currently, traditional 2D graphics technology with low cost and low programming complexity is popularly used in simulating scientific phenomena in school instructions to aid students’ understanding of abstract concepts and principles, while AR technology is relatively new for simulating scientific phenomena. In order to promote the widespread use of AR-supported simulation, it is critical to identify the unique value of AR-supported simulation compared to traditional 2D-supported simulations. So, a 2D-based condition was designed to further compare the influences of AR-supported simulation and 2D-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning. Third, experimenting 39 with scientific simulations is commonly used at the initial stage of teaching new knowledge at school. So the participants were required to have little knowledge about the elastic collision and between-subjects design was applied to compare the impacts of different instructional media on mutual engagement of social interaction in collaborative learning. The whole process of each group’s collaboration was videotaped, and the video recordings served as the primary data source for analyzing the social process of collaborative learning. Both quantitative content analysis and conversation analysis of the social process in collaborative learning were adopted in the data analysis. And the influences of AR-supported simulation on mutual engagement of social interaction were investigated by comparing the equality and mutuality of engagement of social interaction in collaborative learning across three experimental conditions. 3.2 Participants 60 undergraduate students from the National University of Singapore participated in this research. The criterion for being a participant was that he/she must have taken Physics as a subject in Secondary School but not taken it in Junior College/Polytechnic. This ensured that the participants had basic knowledge of motion and energy, but did not know about linear momentum and elastic collision. The sample included 44 females and 16 males, whose age ranged from 21 to 27 years old (M=21.98, SD=1.36). All the participants had no experience of using AR technology before. 10 pairs of participants were assigned to each of the three conditions. 40 3.3 Materials 3.3.1 The systems A mobile AR system was developed to simulate the phenomena of elastic collision in this research. The software prototype was implemented on HTC Nexus One phone running Android OS 2.2 with a supporting server program on a PC. In the system, computational intensive tasks like marker detection and physics simulation had been offloaded to the dedicated server and then the processed results would be sent back to the mobile phone for 3D graphic rendering and display. The physic engine had been built on the server side to detect the collision between the two virtual objects and response to occurring collision according to the principles of physics. Communication between the server and the client were facilitated by high speed Wi-Fi (IEEE802.11) network and the protocols to enable smooth information exchange. This system can visualize two 3D virtual cubes on a marker and has the capacity of simulating the phenomena of elastic collision in a shared virtual space with mobile phones. Each virtual cube is controlled by one user and the user can freely alter the mass and the initial velocity of the cube that he/she controls through the input surface. The simulation process only starts after receiving the data from both users. The whole collision process is visualized with real-time numerical data of mass, velocity, momentum and kinetic energy of the two objects, which are displayed on the two sides of the screen. In the collaboration process, the two users are free to choose when to use the system and they are allowed to run the simulations as many times as they may need to support their discussion. In addition, a simulation supported by 2D graphics technology on the same HTC Nexus 41 One phone was built. To make it a similar architecture to the AR-based simulation, a server was also included in the design. Similarly, each object is controlled by one user and the user can set mass and initial velocity for the object he/she controls. After the p rogram on the server received the information that both users had entered in mobile client s, the simulation would start. Real-time data of mass, velocity, momentum and kinetic energy are also displayed on the two sides of the screen. In the figure 3.1, the view of the AR-supported simulation, the input interface of the AR-based simulation and the view of the 2D-based simulation were presented (from the left to the right). Figure 3.1 The views of the AR-supported simulation, the input interface and the 2D-supported simulation 3.3.2 The task A discussion task related with elastic collision was designed for the two participants in a group to collaboratively work on the problem-solving. The goal of the group discussion was to reach shared understanding on the characteristics of phenomena of elastic collision and the physics principles underlying the mechanism of the phenomena within the group. The questions in the discussion task were as follows: (1) Under the context that the object B is stationary and the object A moves towards B 42 (See Figure 3.2), how many kinds of subsequent motions can happen after the elastic collision? And how does the relationship between the masses of two objects influence the subsequent motions of the two objects after the elastic collision? (2) How do you explain the change of motions of the two objects after elastic collision? Figure 3.2 The learning scenario of the discussion task V1 V2=0 A B m1 m2 3.4 Procedure In the experiment, the two participants within a group were first asked to read a set of paper-based instructional material on elastic collision for 15 minutes independently. The material was extracted from the notes prepared by the physics department of a local Junior College. For the group assigned to the paper-based condition, the questions in the discussion task were introduced to the participants in the group right after the individual reading. And they were to discuss the questions only with the support of the instructional material. For the groups assigned to the AR-based or 2D-based condition, the participants were instructed the method of manipulating the systems after the reading. Each group practiced to use the system for several rounds to be familiar with the way of using it. Next, the discussion task was presented to them and they started collaboratively solving the problems required with the simulation support. In all the three experimental conditions, once the two participants in a 43 group had reached an agreement on the answers to the questions, they would submit a discussion summary and the whole experiment ended. The scenarios of the experiments in the paper-based, the 2D-based and the AR-based conditions (from the left to the right) were showed in Figure 3.3. Figure 3.3 The scenarios of experiments in the three conditions 3.5 Data Analysis The video recordings of collaborative learning of 30 groups’ were transcribed for the analyses of the influences of AR-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning for physics. This research focused on the knowledge-based social interaction in the process of collaborative learning, while procedural interactions for organizing the discussion (e.g. dividing roles for running experiments, managing discussion flow between questions), asking and providing information on setting up the experiments (e.g. initiating a simulation, asking and offering information on setting values before experimenting the simulation) and off-task interactions (e.g. making comments on the technology, expressing personal affairs) were not included in the analyses of mutual engagement of social interaction in collaborative learning. The data analysis was comprised of two stages. At the first stage, the quantitative 44 content analysis based on the coding scheme was used to specify the general patterns and assess the quality of social interaction in collaborative learning across three conditions. At the second stage, the conversation analysis functioned as a more situated approach to further the illustration of the findings of the quantitative content analysis in the conversation contexts and deepen the understanding of opportunities created by AR-supported simulation to support mutual engagement of social interaction in collaborative learning. 3.5.1 Quantitative content analysis 3.5.1.1 Mutual engagement of social interaction Interaction sequences, described as episodes of conversation with the same level of content focus in the social process of collaborative learning, functioned as contexts to examine mutual engagement of social interaction in collaborative learning in this research. In collaborative learning, the interaction sequence was formed when one person in the group initiates a new level of content focus in the conversation and the other person must give a response at the same level of focus. Then a subset of verbal exchanges might take place at the same level of content focus in the group, which in turn led to further development of the interaction sequence. The interaction sequence ended when one person in the group shifted the level of content focus by making a request for new information or presenting information at a new level of content focus instead of directly building on the prior statement. Therefore, each interaction sequence was comprised of an initiating statement that controlled the conversation direction and a series of development statements that extended the initiation at the same level of content focus. According to the conceptualization of mutual engagement of 45 social interaction in peer collaboration proposed by Damon and Phelps (1989), the degree of variation of the number of initiating statements made by two collaborators and the approaches of constructing development statements following the initiating statement were primary indicators of the equality and mutuality of engagement of social interaction in collaborative learning. In this research context, initiating statements referred to presenting a statement or asking a question at a new level of content focus in the process of collaborative learning, while development statements were those following up the initiating statement and extended the discussion centering upon the same content focus. The content focuses could be different aspects related with elastic collision, for example, describing one possible results of elastic collision, interpreting the mechanism underlying the elastic collision, predicting possible results of elastic collision, summarizing the findings, etc. Measures of equality of engagement of social interaction The equality of engagement of social interaction was examined based on the degree of equality on controlling over the direction of conversation for problem-solving. The number of initiating statements made by each participant in a group was counted, and the equality of engagement of social interaction was measured by the standard deviation of the amount of initiating statements engaged by each participant in the social process of collaborative learning to represent the degree of variation between the number of initiating statements expressed by the two collaborators (referring as “equality index”). The higher the standard deviation was, the lower level of the equality was (Jahng, Nielsen, & Chan, 2010). Then, the average equality index of all groups in each condition was calculated. One-way, 46 between-groups ANOVA and post-hoc comparisons were used to test the differences in average equality index across three conditions to identify the impacts of AR-supported simulation on the equality of engagement of social interaction in collaborative learning. Measures of mutuality of engagement of social interaction In order to gain an insight into the characteristics of mutuality of engagement of social interaction in collaborative learning, communicative functions of development statements were analyzed on the basis of the coding scheme developed by Barron (2000) and modified for this research (See Table 3.1). Five mutually exclusive categories of development statements were defined, including elaborations, clarifications, accumulations, acceptances and rejections/no responses. They were analyzed at the level of utterance that was defined as the statement with single communicative function made by one person (Visschers-Pleijers, Dolmans, De Leng, Wolfhagen, & Van der Vleuten, 2006). These categories could facilitate to identify patterns of social interaction with different levels of mutuality, which were beneficial for assessing the effectiveness of AR-supported simulation for supporting mutual engagement of social interaction in face-to-face collaborative learning for physics. Elaborations included developing the prior statement by offering an alternative explanation, disagreeing with the statement and providing logical justifications, modifying the prior statement with reasonable argumentation, or proposing a hypothesis to interpret the prior statement. They were perceived as the development statements with high mutuality to effectively construct mutual understanding. The person did not only positively build upon the prior statement expressed by the partner, but also presented critical viewpoints towards the prior statement for further meaning negotiation. 47 Clarifications referred to responses that requested for more information, explanation or verification in relation to the prior statement at the same level of content focus. They were also characterized by high mutuality. The person needed to actively seek to deepen the mutual understanding after integrating the meaning of the knowledge expressed by the partner with his/her prior understanding. Accumulations were responses that accepted the prior statement and offered additional supportive information as warrant or took up the prior request and provided an answer to it. They played an important role in accumulating common grounds between the collaborators. But compared to elaborations and clarifications, the mutuality of accumulations was lower since less exploratory thinking was involved in the statements. Acceptances included showing simple agreements with the prior statement, such as “Yea.”, “Ok.”, “Yes, I think it’s right.”, or just repeating the prior statement expressed by his/her partner. Although they contributed to reaching agreement rapidly, they were low in meaning richness, which made the development of mutual understanding less constructive. Hence, the mutuality of acceptances was lower than accumulations. Rejections/no response included simply rejecting the prior statement without giving any reasons or ignoring the other’s initiating statement or the prior clarification. The responses that made no substantive contribution to the conversation such as “I don’t know” and “I’m not sure” also fell into this category. Featured by less reciprocity, this type of statements hindered the construction of mutual understanding and was the lowest in mutuality of engagement of social interaction. 48 Table 3.1 Coding scheme for knowledge-based social interaction Category Description Initiated a new level of content focus in the conversation to direct the Initiating statements interaction flow, including requesting for information and presenting a statement at a new level of content focus Development A person’s response to the prior statement at the same level of content statements focus Developed the prior statement by providing an alternative Elaborations explanation; disagreed with or modified the prior statement with rationales; proposed a hypothesis to interpret the prior statement Requested for more information, explanation, verification in relation Clarifications to the prior statement Accepted the prior statement and offer additional information as Accumulations warrant; took up the prior request and provide an answer to it Acceptances Simply agreed with the prior statement; repeat the prior statement Rejected the prior statement without giving any reasons; ignored the Rejections/no prior initiating statement or the clarification within an interaction responses sequence; non-substantive responses Regarding the evaluation of the mutuality of engagement of social interaction in collaborative learning, the occurrences of five categories of development statements within a group were firstly counted. Then, the proportion of each category of development statements in the group was derived by dividing the frequency of each category of development statements by the overall amount of development statements in the social process of collaborative learning. The average percentage of each category of development statements of all groups in each condition was calculated. The assumption of one-way ANOVA was not met 49 when comparing the percentage with different denominators. So, to identify the impacts of AR-supported simulation on the mutuality of engagement of social interaction in collaborative learning, nonparametric Kruskall-Wallis tests were used to test the differences in the average percentage of each category of development statements across three experimental conditions and Mann-Whitney tests were applied for pairwise comparisons. 3.5.1.2 Inter-coder reliability The inter-coder reliability of classifying communicative functions of knowledge-based social interaction was tested in this research. 12 transcripts (40% of all transcripts) were randomly selected from the overall 30 transcripts, and two independent coders identified the communicative functions of knowledge-based social interaction in each transcript according to the coding scheme. The inter-coder reliability (Cohen’s kappa) was 0.793, which showed a substantial agreement. 3.5.2 Conversation analysis Conversation analysis of social interaction in collaborative learning is identified as a useful way to better interpret the reasons that result in the variability of patterns of social interaction across different situations found in the quantitative analysis (Barron, 2003). In this research, in addition to comparing mutual engagement of social interaction on the basis of quantitative content analysis, the conversation analysis of the social process in collaborative learning was applied. Through contextualizing the situations when collaborators initiated a new focus in the conversation and adopted patterns of social interaction with different levels of mutuality, it helped to explore the reasons that related to the differences in the equality and 50 mutuality of engagement of social interaction across three conditions. In the conversation analysis, interaction sequences maintained the contexts for evaluating the influences of AR-supported simulation on mutual engagement of social interaction in face-to-face collaborative learning. Initiating statements and development statements with different levels of mutuality proposed in the quantitative content analysis served as primary discourse events when analyzing sequential evolution of collaborators’ negotiation of meaning. Also, the content of knowledge-based social interaction for constructing shared understanding of elastic collision was examined. Two main categories of content were identified. One was describing the surface phenomena of elastic collision, such as the masses of the two objects and the velocities of the two objects before and after elastic collision. The other was explaining the underlying mechanism underlying the phenomena, mainly including applying the concepts and principles in the instructional material or life experience to illustrate the change of motions before and after the collision. Specifically, the attention was paid to the way that collaborators integrated the experience of manipulating simulations with the social process in the 2D-based and AR-based conditions for building mutual understanding, which included the language used to describe and reflect on the simulation scenarios during the discussion. By situating the occurrences of initiating statements and development statements featured by different levels of mutuality in the context of interaction flow and using simulation support, I attempted to examine how the differences in mutual engagement social interaction emerged during meaning negotiation among three conditions and capture typical examples to interpret the challenges and opportunities that resulted in different levels of 51 mutual engagement of social interaction to build mutual understanding in collaborative learning. 52 Chapter 4 Results The results of data analyses are presented in the following two sections. In the first section, the differences of the equality and mutuality of engagement of social interaction in collaborative learning across three experimental conditions in the quantitative analysis are compared through testing the hypotheses proposed in this research. Based on those significant influences exerted by the AR-supported simulation in the first section, in the second section, representative episodes of conversation in three conditions are analyzed to provide more detailed evidence on how the AR-supported simulation impacts mutual engagement of social interaction in face-to-face collaborative learning for physics. 4.1 Quantitative Analyses of the Influences of AR Technology on Social Interaction In this part, the influences of AR-supported simulation on the equality of engagement of social interaction are quantitatively analyzed at first. Then I focus on the mutuality of engagement of social interaction and compare patterns of social interaction across three experimental conditions. 4.1.1 Equality of engagement of social interaction The equality of engagement of social interaction was measured by the level of equality on controlling over the direction of social interaction flow. Initiating statements indicated the change of direction of social interaction in collaborative learning, and the 53 number of initiating statements made by each person represented his/her control of the direction. Equality index, the standard deviation of the amount of initiating statements generated by each person in a group, was used to assess the degree of variation between the amounts of initiating statements made by two collaborators, which represented the level of equality on controlling over the task. Then the average equality index of groups in each condition was gotten for further comparison across three conditions. The equality index with a bigger number meant that there was a large variation between the amounts of initiating statements made by two collaborators, indicating the level of equality of engagement of social interaction was lower. The average equality index of two students within the groups in the paper-based condition (5.59) was much bigger than those in the 2D-based condition (2.12) and the AR-based condition (1.56) (See Table 4.1). Thus, the equality of engagement of social interaction in the AR-based condition was the highest, which was followed by those in the 2D-based condition and the paper-based condition. The difference in the level of the equality of engagement of social interaction in collaborative learning across three conditions was significant (F (2, 27) =5.733, p[...]... collaborative learning And patterns of social interaction with different levels of mutuality are identified to facilitate the evaluation of mutuality of engagement of social interaction In sum, the objective of this research is to investigate the influences of AR- supported simulation on mutual engagement of social interaction in face- to -face collaborative learning for physics Since equality and mutuality of engagement. .. process of collaborative learning Recognized the significance of joint commitments in effective interpersonal communication, extending it to the context of collaborative learning is useful for analyzing the efficacy of social interaction to maintain mutual understanding of knowledge in AR- supported collaborative learning Mutual engagement of social interaction, emphasizing on joint commitments in social interaction. .. reducing the use of patterns of social interaction with low mutuality are essential to promote the mutuality of engagement of 5 social interaction Thus, to assess the effectiveness of AR- supported simulation in enhancing mutual engagement of social interaction, this study attempts to examine the effects of AR- supported simulation on the equality and mutuality of engagement of social interaction in collaborative. .. understand the efficacy of social interaction in CSCL, two theoretical approaches are adopted to study social interaction in collaborative learning and mutual engagement of social interaction in the social process Next, I review the literature of technology-based simulation on social interaction in collaborative learning Finally, I introduce the potential role of AR- supported simulations in fostering face- to -face. .. understanding Mutual engagement of social interaction, stressing joint commitments in social interaction for reaching mutual understanding, has emerged as a vital relational issue in the research on social interaction of collaborative learning (Damon & Phelps, 1989) In the following part, relevant literatures on mutual engagement of social interaction in collaborative learning are reviewed 2.2.2 Mutual engagement. .. with collaborative learning, understanding the mediating function of technologies in collaborators’ coordination of social process to build mutual understanding is helpful for gaining an insight into the construction of knowledge in CSCL In order to better understand the effects of AR- supported simulation on strengthening social interaction for developing mutual understanding in collaborative learning, ... and mutuality of engagement of social interaction are important to create a constructive collaborative learning context, it is necessary to explore the impact of technologies on mutual engagement of social interaction in face- to -face collaborative learning Tao (1999) has introduced mutual engagement of social interaction into the research of CSCL for science subjects He used equality (high/low) and mutuality... important to promote the mutuality of 23 engagement of social interaction Therefore, increasing critical elaborations, clarifications, accumulations and reducing simple acceptances, rejections and no responses are significant to enhance the mutuality of engagement of social interaction in collaborative learning Extending mutual engagement of social interaction to the research on CSCL is helpful for enriching... technologies influence the quality of social interaction and thus provides limited knowledge on the effectiveness of technologies to foster social interaction in collaborative learning Adopting the features of social interaction that could reflect the quality of social interaction is significant for analyzing the impact of technologies on the efficacy of social interaction in collaborative learning Since the. .. the initiating statement and the development statements in an interaction sequence facilitate to get a deep insight into mutual engagement of social interaction in collaborative learning The equality of engagement of social interaction highlight the relationship of peers formed in the social process of collaborative learning for seeking mutual understanding (Damon & Phelps, 1989) Different levels of ... issue of enhancing mutuality of engagement of social interaction is increasing the use of patterns of social interaction with high mutuality and reducing the use of patterns of social interaction. .. mutuality of engagement of social interaction In sum, the objective of this research is to investigate the influences of AR-supported simulation on mutual engagement of social interaction in face-to-face. .. learning is useful for analyzing the efficacy of social interaction to maintain mutual understanding of knowledge in AR-supported collaborative learning Mutual engagement of social interaction,